Association for Behavior Analysis International

This symposium presents three studies that examine aspects of delivering incentives as a behavior-change intervention. The components of incentive delivery evaluated and discussed in the presentations include: feedback, probability of delivery, magnitude of reward, and the interplay of social influences. Taken together, these presentations will provide practitioners and researchers with the latest information on what variables are likely to influence incentive systems.

In practice, the delivery of consequencesincluding incentivesdoes not always correspond with the ideal conditions for effective reinforcement). The present study was conducted as an extension to previous research focusing on the effects of uncertainty for incentive delivery. Using a simulated work task in an analogue setting, an incentive was arranged for meeting a performance criterion during each trial with varying degrees of probability applied to the criterion. Specifically, participants were told at the beginning of each trial that the probability of the incentive criterion being applied to their performance was either 5%, 10%, 25%, or 90%. If the criterion was not applied, no incentive was delivered regardless of performance during that trial. In a reversal design, two magnitudes of incentives ($0.75 and $1.50) were compared. During the small incentive condition, the probabilistic schedules of incentives failed to sustain responding on the work task for under at least the lower probabilities. For those three participants, increasing the magnitude of the incentive shifted responding with lower probabilities of earning an incentive were more effective maintaining responding than under the small incentive. This effect of magnitude may have implications for the implementation of incentive systems under less than ideal conditions.

Lotteries have proven to be an effective procedure in changing employee behavior. However, lotteries contain multiple components (e.g., drawing locations, odds of winning, delivery of tickets after performance). The degree to which the components influence the effectiveness of the intervention is largely unknown. Recent research has examined probabilities that a lottery will payout to entrants, but no research has yet to evaluate aspects of the lottery drawing. This investigation extends previous research by examining the social aspects of lotteries by comparing private versus public drawings of lottery winners. Participants were direct care employees in three separate classrooms in a school for children diagnosed with autism. The dependent variable was the number of FAST assessments completed by the employees. A multiple baseline with a multielement component was used to evaluate no lottery, lottery with a winner drawn and informed in private, and a lottery where the winner was drawn and announced in the classroom.